Fourier spectrum of halftone images

Abstract

An analytic expression has been developed that can be used for computing the Fourier spectrum of any halftone image as a function of the original continuous-tone image and the halftone process. This result takes into account the nonlinear effects of the threshold step in the halftone process. Although the result is general and can be applied to any original image, in some cases of periodic images the Fourier components of the halftone image are particularly easy to compute. We show important differences between spectra of halftone images and scanned or sampled images. In halftone images, significant image detail is maintained at spatial frequencies greater than half of the screen frequency and aliasing or moiré patterns depend upon object contrast and upon the halftone function differently from similar patterns which occur in scanned or sampled imagery.